146
droplets with realistic compositions, accounting for important thermodynamic and
biochemical processes in both the near and far fields. The dynamic coupling of
TAMOC and oil-CMS uses observed in situ temperature, salinity, pressure, and
velocity profiles in the near-field computations to capture the formation of the deep
intrusion that formed near 1100 m depth. As the live oil rises in the water column,
TAMOC tracks the evolving composition of droplets and passes their mass flow
rates and properties off to the far-field oil-CMS model as initial conditions for the
far-field computations of individual droplets transport and fate, taken from an
assumed droplet size distribution (DSD). TAMOC simulations are described in Gros
et al. (2016, 2017). Here the original 279 pseudo-components were reduced to 19,
including 5 soluble pseudo-components that showed significant extent of aqueous
dissolution during the DWH accident (Reddy et al. 2012; Ryerson et al. 2012), to
reduce the computational cost of simulations. Similarly to the Gros et al. (2017)
paper, the model includes two different scenarios, one with SSDI and one without.
TAMOC output describing distinct droplet types is converted into the individualbased Lagrangian framework of oil-CMS following Eq. 9.1. Because TAMOC considers a finite number of particle sizes and oil-CMS initializes particles from a
continuous size distribution, the discrete droplet diameters from TAMOC are fit to
a CMS probabilistic droplet size distribution (DSD) while ensuring that mass is
conserved. The pseudo-component mixture is represented in oil-CMS by multiple
fractions (pseudo-components) within individual droplet (Lindo-Atichati et al.
2016), and processes affecting oil fate (biodegradation, sedimentation, landfall) are
calculated for each fraction. Droplet movement is a result of their displacement due
to the deterministic velocity field, turbulent mixing (Paris et al. 2013), and buoyancy
(Zheng et al. 2003). For these DWH simulations, environmental variables, such as
horizontal and vertical velocity, temperature, and salinity, are from the Gulf of
Mexico HYCOM hindcast (0.04 degree horizontal resolution, 20 vertical layers).
Values for the ranges of the biodegradation rates for each pseudo-component are
compiled from the list of half-lives of chemical compounds from Prosser et al.
(2016). For pseudo-components with a small number of data points (N < 6), the
minimum and maximum half-lives were used for the faster and slower biodegradation rates, respectively, while for pseudo-components with a larger number of data
points (n > 6), the 25th and 75th percentiles were used.
Different droplet sizes released in the far field at the same time form contrasting
horizontal dispersion and vertical layers (Fig. 9.1), indicating that the initial density
and diameter of droplets are a major driver of their vertical distributions and, ultimately, oil fate. Smaller and denser droplets (Fig. 9.1a), such as the microdroplets
added in the near-field model to account for tip-streaming by dispersant injection,
remain at depth, following the deep intrusion layer which extends over 300 m of the
water column near 1100 m depth. In contrast, larger, less dense droplets ascend into
the water column rapidly (Fig. 9.2b–d). We further observe that oil droplets sequestered in the deep intrusion layer are only formed for the chemically dispersed case
(CD, i.e., including SSDI), which is expected given the large initial DSD from
TAMOC used in our simulations for the naturally dispersed case (ND, i.e., without
SSDI) (Figs. 9.2 and 9.3).
A. C. Vaz et al.
droplets with realistic compositions, accounting for important thermodynamic and
biochemical processes in both the near and far fields. The dynamic coupling of
TAMOC and oil-CMS uses observed in situ temperature, salinity, pressure, and
velocity profiles in the near-field computations to capture the formation of the deep
intrusion that formed near 1100 m depth. As the live oil rises in the water column,
TAMOC tracks the evolving composition of droplets and passes their mass flow
rates and properties off to the far-field oil-CMS model as initial conditions for the
far-field computations of individual droplets transport and fate, taken from an
assumed droplet size distribution (DSD). TAMOC simulations are described in Gros
et al. (2016, 2017). Here the original 279 pseudo-components were reduced to 19,
including 5 soluble pseudo-components that showed significant extent of aqueous
dissolution during the DWH accident (Reddy et al. 2012; Ryerson et al. 2012), to
reduce the computational cost of simulations. Similarly to the Gros et al. (2017)
paper, the model includes two different scenarios, one with SSDI and one without.
TAMOC output describing distinct droplet types is converted into the individualbased Lagrangian framework of oil-CMS following Eq. 9.1. Because TAMOC considers a finite number of particle sizes and oil-CMS initializes particles from a
continuous size distribution, the discrete droplet diameters from TAMOC are fit to
a CMS probabilistic droplet size distribution (DSD) while ensuring that mass is
conserved. The pseudo-component mixture is represented in oil-CMS by multiple
fractions (pseudo-components) within individual droplet (Lindo-Atichati et al.
2016), and processes affecting oil fate (biodegradation, sedimentation, landfall) are
calculated for each fraction. Droplet movement is a result of their displacement due
to the deterministic velocity field, turbulent mixing (Paris et al. 2013), and buoyancy
(Zheng et al. 2003). For these DWH simulations, environmental variables, such as
horizontal and vertical velocity, temperature, and salinity, are from the Gulf of
Mexico HYCOM hindcast (0.04 degree horizontal resolution, 20 vertical layers).
Values for the ranges of the biodegradation rates for each pseudo-component are
compiled from the list of half-lives of chemical compounds from Prosser et al.
(2016). For pseudo-components with a small number of data points (N < 6), the
minimum and maximum half-lives were used for the faster and slower biodegradation rates, respectively, while for pseudo-components with a larger number of data
points (n > 6), the 25th and 75th percentiles were used.
Different droplet sizes released in the far field at the same time form contrasting
horizontal dispersion and vertical layers (Fig. 9.1), indicating that the initial density
and diameter of droplets are a major driver of their vertical distributions and, ultimately, oil fate. Smaller and denser droplets (Fig. 9.1a), such as the microdroplets
added in the near-field model to account for tip-streaming by dispersant injection,
remain at depth, following the deep intrusion layer which extends over 300 m of the
water column near 1100 m depth. In contrast, larger, less dense droplets ascend into
the water column rapidly (Fig. 9.2b–d). We further observe that oil droplets sequestered in the deep intrusion layer are only formed for the chemically dispersed case
(CD, i.e., including SSDI), which is expected given the large initial DSD from
TAMOC used in our simulations for the naturally dispersed case (ND, i.e., without
SSDI) (Figs. 9.2 and 9.3).
A. C. Vaz et al.